Crossover: 2014

Chapter 238 The truth is often in the hands of a few
Even when a large number of schools have not established evaluation indicators suitable for post characteristics, subject characteristics, and research properties.

Things like the number of SCI papers and the impact factor of papers have even become hard currency in the academic circle.

At least at this time in the academic circle (domestic) SCI is quite valuable.

When it comes to future development, Lin Hui can't be said to be obsessed with academics.

In the next few years, major colleges and universities abandoned the "thesis-only theory".

It is required that evaluation paths with different emphases should be established for different types of scientific research work.

The evaluation focuses on the innovation level and scientific value of the paper, and does not take the relevant indicators of SCI papers as the direct basis for judgment;

The focus of technical evaluation for solving key technical problems in production practice should focus on actual contributions (for example, the actual effects of new technologies, new products, and new processes to achieve industrial application)

Papers should not be used as a single basis for evaluation.

After that, SCI was not so deified.

Before that, SCI has always been on the altar.

But things seem to be a little different now.

After rebirth, Lin Hui felt that SCI was nothing more than that, and it wasn't that far away.

Not only is it not far away, but it is very easy.

If the scientific research dog in the next few years knows Lin Hui's emotion at this time, he probably has the intention of killing people.

But it is.

If you have the ability, you will be reborn too...

Anyway, Lin Hui thinks that the paper compiled by Eve Carly only needs to be sent out.

There is absolutely nothing wrong with SCI.

Can articles at the level of general journals in previous lives become SCIs in this time and space?

It seems unbelievable, but it is true.

The progress of the times is fast.

Lin Hui read the papers compiled by Eve Carly.

Although the technology inside is nothing to Lin Hui.

But there are too many new and groundbreaking ideas for people on this plane.

After posting such an article, Lin Hui felt that SCI was absolutely fine.

Even SCI District [-] is not a big problem.

And with the help of some academic channels of MIT.

It is not impossible for the published papers to even go directly to the top journals.

Although Eve Carly has sorted out the results of Lin Hui and his recent discussions in the form of papers, and the sorting is not bad.

But Lin Hui still decided to improve on the basis of the thesis written by Eve Carly.

After all, this is Lin Hui's purely academic debut in this time and space.

For this first show, Lin Hui hopes to achieve perfection.

Although there is almost no perfect thing in the world.

But Lin Hui's philosophy has always been to either not do it, or to do it to the extreme.

Driven by this belief, Lin Hui, the first paper after rebirth, will of course go all out.

Although it is not the first time Lin Hui has written a thesis.

But this kind of thing obviously can't remember.

Although the code is sometimes written irregularly, it may run inexplicably.

But if there are too many loopholes in the paper or something.

That would be ridiculous.

In short, although Lin Hui is in a hurry to produce results in this aspect of the thesis.

But when it comes to specific actions, there is no rush.

It seems that it can only be polished with time.

In short, it was destined to be a sleepless night.

Originally, Lin Hui didn't need to be so anxious.

However, the email sent by Eve Carly also mentioned the follow-up of the generative summary algorithm in the United States.

Although the situation is not pessimistic, it is not very optimistic either.

After Lin Hui tinkered with the generative summary algorithm.

Many commercial scientific research institutions in the United States are rapidly following up the research on forest ash.

In addition, many American universities (including but not limited to Massachusetts Institute of Technology, Stanford University, Carnegie Mellon University, etc.) with strong computer strength are also following this direction.

It is not surprising that these overseas scientific research institutions will quickly follow up the research on Lin Ash.

Although the research related to the subdivision of natural language processing, which is text summarization, does not seem very eye-catching.

Most of the common people don't even know that there are people working on it.

But this does not hinder the significance of text summarization for the progress of human civilization.

Lin Hui has conducted many demonstrations in this regard before.

In fact, the overseas scientific research teams of this time and space should attach great importance to the research of text summarization from the beginning.

It's just that the emphasis has now been raised to another level.

The reason why the level of attention has been increased by one level is inseparable from the movement made by Lin Hui.

After Lin Hui's research results appeared.

At present, the automatic text summarization technology commonly used at home and abroad can be divided into two types according to the different methods of summarization:

Extractive text summarization and generative text summarization.

The method of extractive text summarization is simple to implement, it just extracts existing sentences from the document to form a summary.

Generative text summarization is to use natural language understanding technology to perform grammatical and semantic analysis of the text, fuse information and generate new summary sentences on this basis.

Because Lin Hui just came up with the generative summary algorithm not long ago.

So now the generative summary algorithm is not too widely used except for the application on Nanfeng APP.

On the contrary, the extractive method is widely used due to some historical changes.

But this does not negate the value of generative text summarization.

The academic level has never been overwhelming for the majority.

Truth is often in the hands of a few.

Extractive text summarization can only be regarded as a combinatorial optimization problem in the final analysis.

After all, this falls short in the face of generative text summarization.

Although the two methods of dealing with the same problem (text summarization) seem to be a bit inappropriate.

What is the purpose of human beings in text summarization?

Even what is the purpose of human research in natural language processing?

After all, it is just to better understand natural language and to be able to process natural language more efficiently.

Measured from this perspective, the ability of generative summary algorithms to understand natural language is undoubtedly higher than that of extractive summary algorithms.

Therefore, it is not biased to call the generative text summarization algorithm superior and the extractive text summarization algorithm inferior.

I think these overseas research teams should also have seen that the generative summary algorithm developed by Lin Hui has improved the machine's ability to understand natural language to a higher level.

Only then will the emphasis on the research on generating summaries be further enhanced.

Have to say, the direction is right.

In fact, due to the rise of artificial intelligence in the previous life, artificial intelligence-based generative text summarization has achieved a qualitative leap. After that, generative text summarization has become the main research direction of generative summarization in one fell swoop.

However, it is not enough for some scientific research institutions to invest more in text summarization algorithms, which is not enough for Lin Hui to pay so much attention to it.

(End of this chapter)